Explanations and Causal Judgments are Differentially Sensitive to Covariation and

Mechanism Information
Nadya Vasilyeva (Vasilyeva@Berkeley.Edu)
Tania Lombrozo (Lombrozo@Berkeley.Edu)
Department of Psychology, 3210 Tolman Hall
Berkeley, CA 94720 USA

Abstract
We report four experiments demonstrating that judgments of
explanatory goodness are sensitive both to covariation
evidence and to mechanism information. Compared to
judgments of causal strength, explanatory judgments tend to
be more sensitive to mechanism and less sensitive to
covariation. Judgments of understanding tracked covariation
least closely. We discuss implications of our findings for
theories of explanation, understanding and causal attribution.
Keywords: explanation; covariation; mechanism; causal
strength; understanding

Suppose you work for an art museum, where youre

tasked with tracking museum statistics, including which
visitors visit each gallery, and which visitors make optional
donations to the museum in the extra donation box near the
exit. In organizing your data, you stumble upon a
correlation between two factors: museum visitors who visit
the portrait gallery are much more likely to make an
optional donation than those who do not. How do you make
sense of this relationship? Does the visit to the portrait
gallery explain why some visitors make donations? Are you
persuaded theres a strong causal relationship between
visiting the portrait gallery and making a donation?
In answering these questions, at least two pieces of
additional information may be relevant. First, how strong is
the correlation between visiting the portrait gallery and
making a donation? If the covariation evidence suggests a
perfect association youll likely respond differently from a
case in which the association is weak. Second, is there a
plausible mechanism linking the candidate cause to the
effect? On the face of it the answer may be no, but
suppose you learn of research in social psychology that
exposure to faces (and particularly to eyes) triggers
mechanisms associated with the maintenance of a pro-social
reputation, increasing cooperative behavior (Bateson,
Nettle, & Roberts, 2006). Would this alter your response?
Decades of research on causal learning have pinpointed
both covariation and mechanism information as relevant to
causal claims (e.g., Cheng & Novick, 1990; Koslowski,
1996; Park & Sloman, 2014), with some debate as to their
relative contributions for different causal judgments (Ahn,
Kalish, Medin & Gelman, 1995; Danks, 2005; Newsome,
2003). However, little is known about how these factors
influence judgments of how good an explanation is, or
about whether and how explanatory and causal judgments
diverge with respect to the relative influence of covariation
versus mechanism information. Addressing these questions
is of interest for several reasons.

First, both philosophers and psychologists are interested

in identifying explanatory virtues characteristics that
make for better explanations, such as simplicity, scope, and
a specification of mechanism (e.g., Lipton, 2004). Research
suggests that people find explanations more satisfying when
they are simple and broad (for a review, see Lombrozo,
2012), with additional evidence that explanations are more
likely to be inferred when they are more strongly supported
by probabilistic evidence (Lombrozo, 2007). However, its
unknown whether explanations are also judged better when
they are merely supported by stronger evidence, without
some other explanatory relationship, such as a known causal
mechanism, also in place. The influence of mechanisms on
judgments of explanation goodness is also unknown,
despite many suggestions that explanations and mechanisms
are closely related (e.g., Ahn & Kalish, 2000; Bechtel &
Abrahamsen, 2005; Machamer, Darden, & Craver, 2000).
Second, its important to consider why mechanism
information is valuable in the first place, whether for causal
or explanatory judgments. For starters, mechanism
information could affect the interpretation of covariation
data, making people more confident that a correlation in fact
supports the candidate causal relationship and is not, for
example, the result of a common cause. Once a causal
relationship is established, information about causal
mechanisms will typically support predictions (Douglas,
2009) and interventions (Woodward, 2000): we can predict
who will make donations by knowing whether they visited
the portrait gallery, and we can make people more likely to
donate by increasing their visits to that part of the museum.
Mechanism information can also support broader
generalizations from one case to another. In learning the
mechanism in our museum example, we become better able
to predict whether visiting a sculpture garden will have the
same effect (it should depend on whether the sculptures
have eyes), and on whether the effect might extend to other
museum transactions (such as recycling ones museum
badge versus buying a souvenir). According to accounts that
link the function of explanation to generalization
(Lombrozo & Carey, 2006), one might predict an especially
strong effect of mechanism information on explanation
judgments.
Third, the relationship between causal explanations and
bare statements of the causal relationship they presuppose is
largely unknown. For instance, in explaining museum
donations by appeal to the portrait gallery, are we committing
to any more or less than the claim that visiting the portrait
gallery causally contributes to museum donations?

Identifying factors that differentially influence matched

explanation and causation claims is a good strategy for
beginning to address this question. If explanation claims can
be reduced to the corresponding causal claims, we might
anticipate differences in the absolute value of ratings
assigned to each claim, but ratings for the different claims
should respond similarly to manipulations of covariation
strength and the presence of a mechanism.
For a similar reason, our experiments consider claims
about understanding, e.g., how well people feel they
understand the relationship between visiting the portrait
gallery and making a museum donation. On some accounts,
understanding amounts to a grasp of causes and/or
explanations (e.g., Strevens, 2008), but empirical research
has not considered how judgments of understanding relate
to causal strength or explanation quality.
To investigate these issues, the experiments that follow
manipulate the strength of covariation evidence and the
specification of a mechanism, and elicit judgments about
explanation goodness, causal strength, and understanding.
To preview our results, we find that judgments of causal
strength are more responsive to covariation than either
explanation or understanding judgments, while explanation
judgments are more sensitive to the specification of a full
mechanism than are causal judgments. In the general
discussion we consider the implications of these results for
the issues raised above.

Experiment 1a
Experiment 1 presented participants with two factors that
were selected such that they would not suggest an obvious
causal relationship. Participants received evidence about the
covariation between these factors that suggested no
relationship, a weak relationship, a moderate relationship, or
a strong (deterministic) relationship. We also manipulated
whether they received information about a possible
mechanism.

Method
Participants Four-hundred-and-ninety-two participants
were recruited on Amazon Mechanical Turk in exchange for
$1.45. In all experiments, participation was restricted to
users with an IP address within the United States and an
approval rating of at least 95% based on at least 50 previous
tasks. An additional 217 participants were excluded for
failing a comprehension check for covariation tables (18),
failing a memory check (199), or both (27).
Materials, Design, and Procedure Participants first
completed a practice session in which they were introduced
Table 1. Sample covariation matrices from Experiments 1-2.
Conditions correspond to P = .04, .33, .64 and 1.

None (nearly)

Weak

Moderate

Strong

to covariation tables and received two problems that tested

for comprehension. They were given feedback and requested
to correct wrong responses. Participants who gave up on
comprehension questions without providing the correct
responses were excluded from further analysis.
Next, participants were presented with eight cause-effect
pairs, selected to minimize prior beliefs about their
relationship. Half of the participants were provided with a
hypothetical mechanism connecting the cause and the effect.
Below is sample text from one item:
160 cyclists participated in a large survey. The survey
included many questions. Two of the questions asked: a.
whether or not the cyclist is a woman b. whether or not
the cyclist has ever been hit by a bus at an intersection.
These two things may or may not be related.
No mechanism: In fact, the researchers who designed the
survey didn't have any particular hypotheses about their
relationship.
Full Mechanism: When designing the survey, the
researchers thought that they would be related as follows:
Women are encouraged to obey rules more than men, so
they stop at intersections for red lights more frequently
than men do. This puts them in bus drivers blind spot, so
they get hit by buses more often than men.
Each cause-effect pair was also accompanied by a
covariation table showing nearly no covariation, weak
covariation, moderate covariation, or strong covariation (see
Table 1). Covariation levels rotated through cause-effect
pairs across participants, and each participant saw two
cause-effect pairs for each level of covariation. A small
amount of noise was introduced into the covariation data in
the second set of tables to avoid presenting participants with
identical tables.
Participants were assigned to one of the three judgment
conditions: causal strength, explanatory goodness, or sense
of understanding. Judgment questions were phrased either at
the type or token level.1 Below are sample judgments for the
cyclist item, with token wording in brackets:
[One of the respondents to the survey was LP, who is a
woman. LP was hit by a bus at an intersection.]
Based on the information you have,
Causal strength: do you think there exists a causal
relationship between [LP] being a woman and [LP]
getting hit by a bus at an intersection? No causal
relationship (1) Very strong causal relationship (9)
Explanatory goodness: please rate how good you think
1

In Experiment 1a, participants who were presented with

judgments in the token format gave higher ratings (M=5.65) than
those presented with the type format (M=5.19, F(1,480)=10.94,
p=.001, p2=.022); however, the effect of format was not
significant in Experiment 1b (F(1, 470)=.611, p=.435), and it did
not interact with any other variables in any experiment, so all
reported analyses collapse across this factor.

the following explanation is: Why do some cyclists get hit

by buses at intersections? Because they are women. [Why
was LP hit by a bus at an intersection? Because LP is a
woman.] Very bad explanation (1) Very good
explanation (9)
Sense of understanding: do you feel you understand the
relationship between [LP] being a woman and [LP]
getting hit by a bus at an intersection? Very weak sense of
understanding (1) Very strong sense of understanding (9).
The order of trials was randomized for each participant.
Finally, as a memory check, participants sorted causes from
distractors and matched them with effects; those who made
one or more errors were excluded from further analyses.

Results and Discussion

Are explanation ratings sensitive to covariation and
mechanism information? Explanatory goodness ratings
were subjected to a 4 (covariation: none, weak, moderate,
strong) x 2 (mechanism: none, full) mixed ANOVA. This
revealed a main effect of covariation evidence,
F(3,474)=118.16, p<.001, p2=.428 (all repeated contrasts
ps<.001), with stronger ratings the stronger the covariation:
Mnone=3.01, Mweak=4.79, Mmoderate=5.56, Mstrong=6.43. There
was also a main effect of mechanism, F(1,158)=7.62,
p=.006, p2=.046, with explanations rated as better when a
full mechanism was provided (M=5.29 vs. M=4.61). In
addition, this effect interacted with covariation,
F(3,474)=3.26, p=.021, p2=.020: a full mechanism
significantly increased ratings when the covariation was
absent (Mdiff=1.29, t(158)=4.56, p<.001), but at higher levels
of covariation this effect did not reach significance (weak:
Mdiff=.55, t(158)=1.74, p=.083; moderate: Mdiff=.75,
t(158)=2.32, p=.022; strong: Mdiff=.13, t(158)=.315, p=.753,
Bonferroni-corrected pcrit=.013). Because this interaction
was not significant in subsequent experiments, we are
inclined to attribute this effect in Experiment 1a to random
variation in the data.
Are explanation, causation, and understanding ratings
differentially affected by covariation information? For
each participant we calculated the slope of ratings as a
function of increasing covariation strength, and we
compared mean slopes across the three judgment types in a
one-way ANOVA, revealing a significant effect, F(2,489) =
16.92, p<.001, p2 = .065. As shown in Figure 1, the mean
slope of causal ratings (M=1.41) was higher than the slope
of explanatory goodness ratings (M=1.10, Tukey HSD p=
.010), which was in turn higher than the slope of
understanding ratings (M=.80, p=.013).
Are explanation, causation, and understanding ratings
differentially affected by mechanism information? A 2
(mechanism: none, full) x 3 (judgment: causal strength,
explanatory goodness, sense of understanding) ANOVA on
ratings revealed a main effect of mechanism, F(1,486)=25.57,
p<.001, p2=.050, with higher ratings for a full mechanism
(M=5.77) than no mechanism (M=5.07), as well as a main

Figure 1: Mean covariation slopes as a function of

judgment type in Experiments 1a, 1b and 2. Error bars: 1SE.
effect of judgment, F(2,486)=32.10, p<.001, p2=.117, with
higher ratings for understanding (M=6.23; Tukey HSD
ps<.001) than either causal strength (M=5.14) or
explanation goodness (M=4.95), which did not differ from
each other (p=.500). The interaction was not significant,
F(2,486)=.24, p=.785.

Experiment 1b
Experiment 1a found that explanations were judged better
the stronger the corresponding covariation evidence, and
when a full mechanism was provided. We also found that
explanation judgments were less sensitive to covariation
evidence than were causal judgments, but more sensitive
than understanding judgments. The effect of mechanism did
not differ significantly across judgment types.
In Experiment 1b we tested whether the specification of a
full mechanism was necessary to observe a mechanism
effect, or whether it would suffice to state that some
mechanism connected the two factors. If people suffer from
an illusion of explanatory depth (Rozenblit & Keil, 2002)
and make do with quite skeletal mechanistic understanding
(Keil, 2003), one might anticipate a boost in judgments
from even a mechanism sketch or placeholder, and that this
would be greater for explanation than causal judgments. We
therefore duplicated the structure of Experiment 1a, but
replacing detailed mechanism descriptions with a
mechanism pointer - the statement that the factors in
question are related via some unspecified mechanism.

Method
Participants Four-hundred-and-eighty-two participants
were recruited on Amazon Mechanical Turk in exchange for
$1.45. An additional 198 participants were excluded for
failing a comprehension check for covariation tables (17),
failing a memory check (181), or both (27).
Materials, Design, and Procedure were the same as in
Experiment 1a, with the exception of the mechanism
statement: the full mechanism was replaced with a general
statement that there exists some multi-step pathway
connecting the cause to the effect, omitting all other details.
Mechanism pointer: When designing the survey, the
researchers thought they would be related by a multi-step
pathway connecting being a woman to being hit by a bus

at an intersection: Women and men behave differently,

and the differences in their behavior on the road result in
a different probability of getting hit by a bus at an
intersection.

Results and Discussion

Are explanation ratings sensitive to covariation and
mechanism information? Explanatory goodness ratings
were subjected to a 4 (covariation: none, weak, moderate,
strong) x 2 (mechanism: none, pointer) mixed ANOVA.
This
revealed
a
main
effect
of
covariation,
F(3,516)=146.50, p<.001, p2 = .460, with higher ratings the
stronger
the
evidence:
Mnone=2.43,
Mweak=4.27,
Mmoderate=4.91, Mstrong=6.01, all repeated contrasts ps<.001).
The main effect of mechanism did not reach significance,
F(1,172)=2.71, p=.102, although the difference was in the
predicted direction: no mechanism M=4.33, mechanism
pointer M=4.73. The interaction was not significant,
F(3,516)=1.09, p=.352.
Are explanation, causation, and understanding ratings
differentially affected by covariation information? As in
Experiment 1a, covariation slopes were analyzed as a
function of judgment in a one-way ANOVA, revealing a
significant effect, F(2,479)=20.41, p<.001, p2=.079. As
shown in Figure 1, the ordering of mean slopes mirrored
Experiment 1a, but the difference between the slopes of
causal (M=1.37) and explanatory (M=1.15) judgments did
not reach significance (Tukey HSD p=.111). The slope for
understanding ratings (M=.67) was significantly lower than
that for causal or explanatory ratings (ps<.001).
Are explanation, causation, and understanding ratings
differentially affected by mechanism information? A 2
(mechanism: none, pointer) x 3 (judgment: causal strength,
explanatory goodness, sense of understanding) ANOVA on
ratings revealed that providing a mechanism pointer did not
significantly raise ratings, F(1,476)=1.82, p=.178: Mnone=5.03
versus Mpointer=5.26, suggesting that a skeletal mechanism
is insufficient to affect judgments. There was again a main
effect of judgment, F(2,476)=32.86, p<.001, p2=.121
(Mcaus=5.06, Mexpl=4.52, and Mund=5.94, all different, Tukey
HSD ps.007) and no interaction, F(2,476)=1.03, p=.360.

Experiment 2
Although providing detailed mechanisms in Experiment 1a
boosted all ratings, the effect was weaker than we expected,
which could have masked differences across judgments. In
particular, it is possible that by presenting Experiments 1a
and 1b as studies about the way people understand data
tables, taking participants through an extensive practice
session focusing on covariation tables, and manipulating
covariation within subjects (while judgment and mechanism
varied between subjects) we artificially drew attention to the
covariation manipulation at the expense of the mechanism
information. To address these concerns, we conducted
Experiment 2, in which we minimized task features that
drew attention to the covariation tables, hoping that it would

set an even playing field for covariation and mechanism

manipulations. We also combined the mechanism
manipulations from Experiments 1a and 1b into a single
variable with three levels (full mechanism, mechanism
pointer, and no mechanism) and manipulated it within
subjects, along with two levels of covariation (none, strong).

Method
Participants Two-hundred-and-fifty-one participants were
recruited on Amazon Mechanical Turk in exchange for
$1.55. An additional 81 participants were excluded for
failing a memory check.
Materials, Design and Procedure Mechanism information
(none, pointer, full) and covariation strength (none, strong)
were manipulated within subjects, and rotated through items
across participants. The type of judgment (explanation
goodness, causal strength, sense of understanding) was
manipulated between subjects.
The materials and procedure were the same as in
Experiments 1a and 1b, with the following exceptions: the
number of items (cause-effect pairs) was reduced to 6 and
the practice session was shortened, as the comprehension
questions about covariation tables were removed to avoid
pragmatic cues that covariation evidence should be
prioritized over mechanism information during the task. All
questions were presented in the token format.

Results and Discussion

Are explanation ratings sensitive to covariation and
mechanism information? Explanatory goodness ratings
were subjected to a 2 (covariation: none, strong) x 3
(mechanism: none, pointer, full) repeated-measures
ANOVA. This revealed a main effect of covariation,
F(1,85)=77.69, p<.001, p2=.478, with higher ratings for
strong covariation (M=5.74) than no covariation (M=2.76).
There was also a main effect of mechanism, F(2,170) =
15.71, p<.001, p2=.156. Repeated contrasts indicated that
ratings increased significantly from no mechanism
(M=3.69) to a mechanism pointer (M=4.34) to a full
mechanism (M=4.72), all ps<.05. The effects of mechanism
and covariation did not interact, F(2,170)=.341, p=.712.
Are explanation, causation, and understanding ratings
differentially affected by covariation information? As in
Experiment 1a, covariation slopes were analyzed as a
function of judgment in a one-way ANOVA, revealing a
significant effect, F(2,248)=21.27, p<.001, p2=.146. As
shown in Figure 1, the ordering of mean slopes was the
same as in Experiments 1a and 1b. The covariation slope for
causal strength ratings (M=4.87) was significantly higher
than the slopes for explanatory goodness (M=2.97) and
understanding ratings (M=2.12, Tukey HSD ps<.001); the
difference between the latter two was not significant (p=.115).
Are explanation, causation, and understanding ratings
differentially affected by mechanism information? A
3 (mechanism: none, pointer, full) x 3 (judgment: causal

Method
Participants Ninety-one participants were recruited on
Amazon Mechanical Turk in exchange for $1.00. An
additional 16 participants were excluded for failing a
memory check.

Figure 2: Mean ratings as a function of mechanism and

judgment type in Experiment 2. Error bars: 1SE.
strength, explanatory goodness, sense of understanding)
mixed ANOVA on ratings showed a significant main effect
of mechanism, F(2,496)=29.39, p<.001, p2=.106. Repeated
contrasts showed that ratings increased significantly from
no mechanism (M=4.54) to a mechanism pointer (M=4.99,
p<.001) to a full mechanism (M=5.34, p=.001). Ratings
were
also
significantly
affected
by
judgment,
F(2,249)=18.67, p<.001, p2=.131: all judgments were
significantly different from each other (Mcaus=4.25,
Mexpl=5.02, Mund=45.65, Tukey ps.022). Although the
interaction did not reach significance, F(2,496)=1.64,
p=.162, the pattern of means in Figure 2 suggested that
providing the full mechanism had the most pronounced
effect on explanation. This was confirmed by a significant
ANOVA on full vs. no-mechanism difference scores,
F(2,248)=3.22, p=.042, p2=.025: a full mechanism
produced a larger boost in explanation ratings than causal
ratings (Mdiff 1.04 vs. .44, Tukey p=.043); understanding
received an intermediate boost (Mdiff=.91, ps.143). In
contrast, the difference between the pointer and nomechanism did not vary across judgments, F(2,248)=1.35,
p=.262. This pattern is also consistent with Experiment 3
which finds that explanation goodness ratings are
significantly more responsive to full mechanism information
than causal strength ratings.

Experiment 3
Focusing on explanation ratings versus causal strength
ratings and on the contrast between no mechanism and a full
mechanism, Experiment 2 produced a double dissociation,
with explanation ratings more sensitive than causal ratings
when it came to mechanisms, and causal judgments more
sensitive than explanation judgments when it came to
covariation. While the differential effect of covariation was
also found in Experiments 1a and 1b, the effect of
mechanism information was not. We therefore sought to
replicate the interactions between mechanism and judgment
in Experiment 2 before drawing strong conclusions. We also
tied the mechanism more closely to each judgment by
embedding the mechanism information in the body of the
explanation and causation statements themselves.

Materials, Design and Procedure Experiment 3 included

the following changes from Experiment 2: the mechanism
information was included in the body of the explanation or
causal statement (e.g., explanation with a mechanism
pointer: MP was hit by a bus at an intersection because MP
is a woman, and there exists a multi-step pathway that
connects being a woman to being hit by a bus: women and
men behave differently, and the differences in their behavior
on the road result in a different probability of getting hit by
a bus at an intersection.); the covariation variable was
dropped; the understanding judgment was dropped; and
both judgment type (causal strength, explanation goodness)
and mechanism (none, pointer, full) were manipulated
within subjects. Judgments were blocked, with the order of
blocks randomized across participants. Prior to the second
block, participants were invited to pay attention to the
changed rating scale. Mechanism levels were randomized
within each judgment block. Items rotated through
conditions across participants.

Results and Discussion

A 3 (mechanism: none, pointer, full) x 2 (judgment:
explanation goodness, causal strength) repeated-measures
ANOVA revealed a significant main effect of mechanism,
F(2,180)=48.71, p<.001, p2=.351, with ratings increasing
from no mechanism (M=1.71) to a mechanism pointer
(M=2.20) to a full mechanism (M=3.43, repeated contrasts
ps.001), and no main effect of judgment, F(1,90)=.99,
p=.323. Critically, there was also a significant interaction
between mechanism and judgment, F(2,180)=3.06, p=.049,
p2=.033. As shown in Figure 3, the differences across
mechanism conditions were more pronounced for explanatory than causal judgments. As in Experiment 2, this
interaction was driven by the difference between the no
mechanism and full mechanism conditions: the comparison
of full minus no-mechanism difference for explanation vs.
casual ratings was significant, t(90)=2.18, p=.032, but the
pointer vs. no-mechanism difference did not vary across
judgments (t(90)=.04, p=.971).

Figure 3: Mean ratings as a function of mechanism and

judgment type in Experiment 3. Error bars: 1SE.

General Discussion
In four experiments we demonstrate that judgments of
explanation goodness are sensitive to both covariation
evidence and mechanism information. Comparing
explanation to other judgments, we observed a consistent
dissociation: explanation judgments were less responsive to
the degree of covariation in the data than were causal
judgments. In contrast, specifying a full mechanism had a
stronger effect on explanations than on causal judgments in
Experiments 2 and 3, which drew less attention to the
covariation tables. Of the three judgment types, sense of
understanding was least responsive to covariation. Overall,
our results indicate that these three types of judgments differ
systematically when it comes to the role of covariation data
and the effects of specifying a full mechanism.
Returning to the issues raised in the introduction, our
findings support some tentative conclusions and raise
additional questions for further study. First, we find that
explanations are judged better when supported by stronger
covariation evidence or by the specification of a
mechanism, and that the benefits of stronger evidence are
not limited to cases in which a mechanism is also specified.
It would be interesting to know whether these two factors
affect explanation ratings for different reasons for
example, covariation might be valuable for purely evidential
reasons, while the specification of a mechanism could be a
genuine virtue in addition to having evidential import.
Second, full mechanism information does appear to have
a larger effect on explanation goodness ratings relative to
causal strength ratings, as might be expected on the view
that explanations are especially geared towards generalization (Lombrozo & Carey, 2006), which full mechanism
information supports. More speculatively, it could also be
that reduced sensitivity to covariation emerges for a similar
reason: a certain degree of resistance to over-fitting the data
from a single sample could help achieve more reliable
generalizations (and indeed, Williams, Lombrozo, &
Rehder, 2013 show that explanation encourages a search for
broad patterns despite inconsistent data).
Third, our findings suggest that explanatory goodness
cannot be reduced, in any straightforward way, to judgments
of causal strength. Similarly, ratings of understanding
diverge from those of either explanation or causation. Our
findings thus call for caution when characterizing one of
these judgments in terms of another, and also raise questions
about the extent to which different kinds of explanatory and
causal judgments could diverge. For instance, evaluating
explanatory goodness could diverge from evaluations of
explanation probability, and evaluations of causal structure
could diverge from those of strength.
In sum, we demonstrate that judgments of causal strength,
explanatory goodness and, to some extent, understanding
respond differently to covariation and full mechanism
information. Explanations surpass causal judgments in their
sensitivity to a full mechanism, and the pattern is reversed
for covariation. Our results present a challenge for proposals
that characterize explanations as identifying causes, and

characterize understanding in terms of grasping causal

relationships and/or explanations. More importantly, these
patterns of divergence can begin to help us understand the
different roles of these judgments in our cognitive lives.

Acknowledgments
This work was supported by the Varieties of Understanding
Project, funded by the John Templeton Foundation.

References
Ahn, W.-K., & Kalish, C. (2002). The role of mechanism
beliefs in causal reasoning. In F. C. Keil & R. A. Wilson
(Eds.), Explanation and cognition. MIT Press.
Ahn, W. K., Kalish, C. W., Medin, D. L., & Gelman, S. (1995).
The role of covariation versus mechanism information in
causal attribution. Cognition, 54, 299352.
Bateson, M., Nettle, D., & Roberts, G. (2006). Cues of being
watched enhance cooperation in a real-world setting. Biology
Letters, 2, 412414.
Bechtel, W., & Abrahamsen, A. (2005). Explanation: a
mechanist alternative. Studies in History and Philosophy of
Biological and Biomedical Sciences, 36(2), 421441.
Cheng, P. W., & Novick, L. R. (1990). A probabilistic contrast
model of causal induction. Journal of Personality and Social
Psychology, 58, 545-567.
Danks, D. (2005). The supposed competition between theories
of human causal inference. Philosophical Psychology, 18,
259272.
Douglas, H. E. (2009). Reintroducing prediction to explanation.
Philosophy of Science, 76(4), 444-463.
Keil, F.C. (2003). Folkscience: Coarse interpretations of a
complex reality. Trends in Cognitive Science, 7, 368-373.
Koslowski, B. (1996). Theory and Evidence. MIT Press.
Lipton, P. (2004). Inference to the best explanation.
Psychology Press.
Lombrozo, T. (2007). Simplicity and probability in causal
explanation. Cognitive Psychology, 55(3), 232257.
Lombrozo, T. (2012). Explanation and abductive inference.
Oxford Handbook of Thinking and Reasoning, 260276.
Lombrozo, T., & Carey, S. (2006). Functional explanation and
the function of explanation. Cognition, 99(2), 167204.
Machamer, P., Darden, L., & Craver, C. F. (2000). Thinking
About Mechanisms. Philosophy of Science, 67(1), 125.
Newsome, G. L. (2003). The debate between current versions
of covariation and mechanism approaches to causal
inference. Philosophical Psychology, 16(1), 87107.
Park, J., & Sloman, S. (2014). Causal explanation in the face of
contradiction. Memory & Cognition, 42(5), 80620.
Rozenblit, L., & Keil, F. (2002). The misunderstood limits of
folk science: An illusion of explanatory depth. Cognitive
Science, 26, 521562.
Strevens, M. (2008). Depth: an account of scientific
explanation. Harvard University Press.
Williams, J. J., Lombrozo, T., & Rehder, B. (2013). The
hazards of explanation: overgeneralization in the face of
exceptions. Journal of Experimental Psychology: General,
142(4), 1006-1014.
Woodward, J. (2000). Explanation and invariance in the special
sciences. The British Journal for the Philosophy of Science,
51, 197 254.